changes mainly in the lexicon, for months of the year, to invoke time…

…-related lexical entries. A bit better than what was there before, though still lacking a CARG value.
delph-in · May 29, 2024 · 4fe97f8 · 4fe97f8
1 parent c5b83ec
commit 4fe97f8
Show file tree

Hide file tree

Showing 9 changed files with 62 additions and 55 deletions.
diff --git a/irtypes.tdl b/irtypes.tdl
@@ -232,7 +232,6 @@ fem_pl_ilr := fem_ilr & pl_ilr.
 neut_sg_ilr := neut_ilr & sg_ilr.
 neut_pl_ilr := neut_ilr & sg_ilr.
 
-
 ; -- determiners
 
 d_ilr :=  infl-ltow-rule & 

diff --git a/letypes.tdl b/letypes.tdl
@@ -1229,7 +1229,6 @@ nom-subj-synsem := lex-synsem &
 
 """.
 
-
 arg1_lt := nom-subj-synsem & 
   [ LOCAL.CAT.VAL.SUBJ < [ LOCAL.CONT.HOOK.INDEX #ind ] >,
     LKEYS.KEYREL arg1-ev-relation & 
@@ -4612,6 +4611,7 @@ v_ap_ser_synsem := v_copula_synsem &
   """
   e.g. Juan es feo
   removed LOCAL [ CAT.VAL COMPS < [ LOCAL [ CAT.HEAD [ KEYS.KEY basic_adj_rel ], to cover "será complicado"
+  OZ 2024-01-31: Added AGR #sind in MOD to rule out "Ella es bueno"
   """.
 
 

diff --git a/lexicon.tdl b/lexicon.tdl
@@ -1094,10 +1094,9 @@ abrillantar_v-np := v_np_native_le &
   [ STEM < "abrillantar" >,
     SYNSEM.LKEYS.KEYREL.PRED "_abrillantar_v_rel" ].
 
-abril_n := n_-_c_native_le &
+abril_n := n_-_c-tmp_native_le &
   [ STEM < "abril" >,
-    SYNSEM.LKEYS.KEYREL [ ARG0.SORT tmp,
-                          PRED "_abril_n_rel" ] ].
+    SYNSEM.LKEYS.KEYREL.ARG0.SORT tmp ].
 
 abrir_v-np-pp_a-sbj_cp_p := v_np-ppa*_sbj-cp-p-sub_native_le &
   [ STEM < "abrir" >,
@@ -8548,10 +8547,9 @@ agostar_v-np_rfx := v_np_rfx_native_le &
   [ STEM < "agostar" >,
     SYNSEM.LKEYS.KEYREL.PRED "_agostar_v_rel" ].
 
-agosto_n := n_-_c_native_le &
+agosto_n := n_-_c-tmp_native_le &
   [ STEM < "agosto" >,
-    SYNSEM.LKEYS.KEYREL [ ARG0.SORT tmp,
-                          PRED "_agosto_n_rel" ] ].
+    SYNSEM.LKEYS.KEYREL.ARG0.SORT tmp ].
 
 agotable_aj-nspd := aj_-_i-nspd_native_le &
   [ STEM < "agotable" >,
@@ -77901,10 +77899,9 @@ diciclopentadieno_n := n_-_nc_native_le &
     SYNSEM.LKEYS.KEYREL [ ARG0.SORT abs,
                           PRED "_diciclopentadieno_n_rel" ] ].
 
-diciembre_n := n_-_c_native_le &
+diciembre_n := n_-_c-tmp_native_le &
   [ STEM < "diciembre" >,
-    SYNSEM.LKEYS.KEYREL [ ARG0.SORT tmp,
-                          PRED "_diciembre_n_rel" ] ].
+    SYNSEM.LKEYS.KEYREL.ARG0.SORT tmp ].
 
 dicloroacetato_n := n_-_c_native_le &
   [ STEM < "dicloroacetato" >,
@@ -88926,10 +88923,9 @@ energúmeno_aj-i := aj_-_i_native_le &
   [ STEM < "energúmeno" >,
     SYNSEM.LKEYS.KEYREL.PRED "_energúmeno_a_rel" ].
 
-enero_n := n_-_c_native_le &
+enero_n := n_-_c-tmp_native_le &
   [ STEM < "enero" >,
-    SYNSEM.LKEYS.KEYREL [ ARG0.SORT tmp,
-                          PRED "_enero_n_rel" ] ].
+    SYNSEM.LKEYS.KEYREL.ARG0.SORT tmp ].
 
 enervar_v-np_rfx := v_np_rfx_native_le &
   [ STEM < "enervar" >,
@@ -104796,10 +104792,9 @@ fealdad_n-pp := n_pp_mc_native_le &
     SYNSEM.LKEYS.KEYREL [ ARG0.SORT abs,
                           PRED "_fealdad_n_rel" ] ].
 
-febrero_n := n_-_c_native_le &
+febrero_n := n_-_c-tmp_native_le &
   [ STEM < "febrero" >,
-    SYNSEM.LKEYS.KEYREL [ ARG0.SORT tmp,
-                          PRED "_febrero_n_rel" ] ].
+    SYNSEM.LKEYS.KEYREL.ARG0.SORT tmp ].
 
 febrífugo_aj-i-nprd := aj_-_i-nprd_native_le &
   [ STEM < "febrífugo" >,
@@ -139880,10 +139875,9 @@ juicioso_aj-i := aj_-_i_native_le &
   [ STEM < "juicioso" >,
     SYNSEM.LKEYS.KEYREL.PRED "_juicioso_a_rel" ].
 
-julio_n := n_-_c_native_le &
+julio_n := n_-_c-tmp_native_le &
   [ STEM < "julio" >,
-    SYNSEM.LKEYS.KEYREL [ ARG0.SORT tmp_unit,
-                          PRED "_julio_n_rel" ] ].
+    SYNSEM.LKEYS.KEYREL.ARG0.SORT tmp ].
 
 julio_n-part := n_pp_psd-part_native_le &
   [ STEM < "julio" >,
@@ -139904,10 +139898,9 @@ jungla_n := n_-_c_native_le &
     SYNSEM.LKEYS.KEYREL [ ARG0.SORT loc,
                           PRED "_jungla_n_rel" ] ].
 
-junio_n := n_-_c_native_le &
+junio_n := n_-_c-tmp_native_le &
   [ STEM < "junio" >,
-    SYNSEM.LKEYS.KEYREL [ ARG0.SORT tmp,
-                          PRED "_junio_n_rel" ] ].
+    SYNSEM.LKEYS.KEYREL.ARG0.SORT tmp ].
 
 junior_n := n_-_c_native_le &
   [ STEM < "junior" >,
@@ -150917,10 +150910,9 @@ marxista_n := n_-_c_native_le &
     SYNSEM.LKEYS.KEYREL [ ARG0.SORT hum,
                           PRED "_marxista_n_rel" ] ].
 
-marzo_n := n_-_c_native_le &
+marzo_n := n_-_c-tmp_native_le &
   [ STEM < "marzo" >,
-    SYNSEM.LKEYS.KEYREL [ ARG0.SORT tmp,
-                          PRED "_marzo_n_rel" ] ].
+    SYNSEM.LKEYS.KEYREL.ARG0.SORT tmp ].
 
 masacrar_v-np := v_np_native_le &
   [ STEM < "masacrar" >,
@@ -151823,10 +151815,9 @@ mayonesa_n := n_-_c_native_le &
     SYNSEM.LKEYS.KEYREL [ ARG0.SORT cnc,
                           PRED "_mayonesa_n_rel" ] ].
 
-mayo_n := n_-_c_native_le &
+mayo_n := n_-_c-tmp_native_le &
   [ STEM < "mayo" >,
-    SYNSEM.LKEYS.KEYREL [ ARG0.SORT tmp,
-                          PRED "_mayo_n_rel" ] ].
+    SYNSEM.LKEYS.KEYREL.ARG0.SORT tmp ].
 
 mayor_aj-pp-cmp := aj_pp_i-cmp_native_le &
   [ STEM < "mayor" >,
@@ -166210,10 +166201,9 @@ novicio_n := n_-_c_native_le &
     SYNSEM.LKEYS.KEYREL [ ARG0.SORT hum,
                           PRED "_novicio_n_rel" ] ].
 
-noviemtre_n := n_-_c_native_le &
-  [ STEM < "noviemtre" >,
-    SYNSEM.LKEYS.KEYREL [ ARG0.SORT tmp,
-                          PRED "_noviemtre_n_rel" ] ].
+noviembre_n := n_-_c-tmp_native_le &
+  [ STEM < "noviembre" >,
+    SYNSEM.LKEYS.KEYREL.ARG0.SORT tmp ].
 
 novillo_n := n_-_c_native_le &
   [ STEM < "novillo" >,
@@ -167726,10 +167716,9 @@ octovalente_aj-i-nprd := aj_-_i-nprd_native_le &
   [ STEM < "octovalente" >,
     SYNSEM.LKEYS.KEYREL.PRED "_octovalente_a_rel" ].
 
-octubre_n := n_-_c_native_le &
+octubre_n := n_-_c-tmp_native_le &
   [ STEM < "octubre" >,
-    SYNSEM.LKEYS.KEYREL [ ARG0.SORT tmp,
-                          PRED "_octubre_n_rel" ] ].
+    SYNSEM.LKEYS.KEYREL.ARG0.SORT tmp ].
 
 óctuple_aj-i-nprd := aj_-_i-nprd_native_le &
   [ STEM < "óctuple" >,
@@ -214632,10 +214621,9 @@ séptico_aj-i-nprd := aj_-_i-nprd_native_le &
   [ STEM < "séptico" >,
     SYNSEM.LKEYS.KEYREL.PRED "_séptico_a_rel" ].
 
-septiembre_n := n_-_c_native_le &
+septiembre_n := n_-_c-tmp_native_le &
   [ STEM < "septiembre" >,
-    SYNSEM.LKEYS.KEYREL [ ARG0.SORT tmp,
-                          PRED "_septiembre_n_rel" ] ].
+    SYNSEM.LKEYS.KEYREL.ARG0.SORT tmp ].
 
 septisómico_aj-i-nprd := aj_-_i-nprd_native_le &
   [ STEM < "septisómico" >,

diff --git a/tsdb/skeletons/mrs/item b/tsdb/skeletons/mrs/item
@@ -31,7 +31,7 @@
 311@unknown@formal@none@1@S@El perro que perseguía a Núria lladró.@@@@1@7@The dog that Browne chased barked.@oe@21-9-2008
 321@unknown@formal@none@1@S@El perro a perseguir está ladrando@@@@1@6@The dog to chase is barking.@oe@21-9-2008
 331@unknown@formal@none@1@S@El perro fue perseguido por Núria@@@@1@6@The dog was chased by Browne.@oe@21-9-2008
-341@unknown@formal@none@1@S@El perro perseguido por Núria ladró@@@@1@6@The dog chased by Browne barked.@oe@21-9-2008
+341@unknown@formal@none@1@S@El perro perseguido por el gato ladró@@@@1@6@The dog chased by the cat barked.@oe@21-9-2008
 351@unknown@formal@none@1@S@El perro está ladrando.@@@@1@4@The dog is barking.@oe@21-9-2007
 361@unknown@formal@none@1@S@El perro ha ladrado.@@@@1@4@The dog has barked.@oe@21-9-2007
 371@unknown@formal@none@1@S@El perro ha estado ladrando.@@@@1@5@The dog has been barking.@oe@21-9-2007

diff --git a/util/freeling_api/srg-freeling.dat b/util/freeling_api/srg-freeling.dat
@@ -37,9 +37,8 @@ alguna @any
 algunos @any
 algunas @any
 rueda @any
-peor @any
-mejor @any
 divertido @any
+pasado @any
 </NoDisambiguate>
 
 ## List of words for which the list of output analysis given
@@ -106,6 +105,7 @@ escasa escaso AQ0FS00 escaso DI0FS0
 escasas escaso AQ0FP00 escaso DI0FP0
 escaso escaso AQ0MS00 escaso DI0MS0
 escasos escaso AQ0MP00 escaso DI0MP0
+helado helado NCMS000 helado VMP00SM
 numerosa numeroso AQ0FS00 numeroso DI0FS0
 numerosas numeroso AQ0FP00 numeroso DI0FP0
 numeroso numeroso AQ0MS00 numeroso DI0MS0
@@ -117,11 +117,24 @@ raros raro AQ0MP00 raro DI0MP0
 cientos ciento Zd
 millares millar Zd
 miles mil Zd
+mejor mejor AQ0CS00 mejor RG
+mejores mejor AQ0CP00
+peor peor AQ0CS00 peor RG
+peores peor AQ0CP00
+mayor mayor AQ0CS00
+mayores mayor AQ0CP00
+menor menor AQ0CS00
+menores menor AQ0CP00
 off-line off-line AQ0CN00
 on-line on-line AQ0CN00
 no_sólo no_sólo CC no_sólo RG
 hace hace SP hacer VMIP3S0
 hacía hacía SP hacer VMII3S0
+cerca cerca RG cercar VMIP3S0
+mar. marzo NCMS000 mar NCMS000
+favorito favorito AQ0MS00 favorito NCMS000
+favoritos favorito AQ0MP00 favorito NCMP000
+
 </ReplaceAll>
 
 ## List of tag fusions to perform. 

diff --git a/util/srg_freeling2yy.py b/util/srg_freeling2yy.py
@@ -40,8 +40,8 @@ def override_tag(selected, word, lemma, tag, override_dicts):
                     return {'tags': [tags[0]+'+'+tags[1][:-3]], 'prob': -1 }
                 else:
                     print("More than four tags in Freeling output: {}".format(selected['tag']))
-    if lemma in override_dicts['replace'] and len(override_dicts['replace'][lemma]['lemma']) == 1:
-        return {'tags': override_dicts['replace'][lemma]['tag'], 'prob': -1 }
+    if word in override_dicts['replace'] and len(override_dicts['replace'][word]['lemma']) == 1:
+        return {'tags': override_dicts['replace'][word]['tag'], 'prob': -1 }
     return {'tags': [selected['tag']], 'prob': selected['prob']}
     #raise Exception("selected tag not in tag list")
 
@@ -77,8 +77,8 @@ def convert_sentences(sentences, override_dicts):
             for j,tok in enumerate(sent['tokens']):
                 is_additional = tok['additional']
                 surface = tok['form']
-                #if surface == 'primer':
-                #    print('debug')
+                if surface == 'mar.':
+                    print('debug')
                 tag_prob = {'tag': tok['tag'], 'prob':tok['prob']}
                 pos_conf = override_tag(tag_prob, surface.lower(), tok['lemma'], tok['tag'], override_dicts)
                 if len(pos_conf['tags']) > 1:

diff --git a/util/tokenize_and_tag.py b/util/tokenize_and_tag.py
@@ -99,7 +99,8 @@ def tokenize_and_tag(self, sentence_list, override_dicts):
                     tag = '" "+'.join([tp['tag'] for tp in tags_probs])
                     prob = tags_probs[-1]['prob']
                     #print("lemma: {}, form: {}, start: {}, end: {}, tag: {}".format(w.get_lemma(), w.get_form(), w.get_span_start(), w.get_span_finish(), w.get_tag()))
-                    output[i]['tokens'].append({'lemma':w.get_lemma(), 'form': w.get_form(),
+                    lemma = w.get_lemma() if not 'lemma' in tags_probs[-1] else tags_probs[-1]['lemma']
+                    output[i]['tokens'].append({'lemma': lemma, 'form': w.get_form(),
                                                 'start':w.get_span_start(), 'end': w.get_span_finish(),
                                                 'tag': tag, 'prob': prob, 'additional': additional})
                     for k,arc in enumerate(additional_arcs):
@@ -129,24 +130,30 @@ def freeling_analyze(self, lin, sid):
     def get_selected_tags(self, w, override_dicts):
         tags = []
         additional_arcs = []
-        #if w.get_form().lower() == "primer":
-        #    print("debug")
+        seen = set()
+        if w.get_form().lower() == "mar.":
+            print("debug")
         for a in w:
             if a.is_selected():
                 if a.is_retokenizable():
                     tks = a.get_retokenizable()
                     for tk in tks:
                         tags.append(({'tag': tk.get_tag(), 'prob': a.get_prob()}))
                 else:
-                    if not w.get_form().lower() in override_dicts['replace']:
+                    needs_replacement = w.get_form().lower() in override_dicts['replace']
+                    if not needs_replacement:
                         tags.append(({'additional':False, 'tag': a.get_tag(), 'prob': a.get_prob()}))
                     else:
                         for i, additional_tag in enumerate(override_dicts['replace'][w.get_form().lower()]['tag']):
                             additional_lemma = override_dicts['replace'][w.get_form().lower()]['lemma'][i]
                             if i == 0:
-                                tags.append(({'additional':True, 'tag': additional_tag, 'prob': -1, 'lemma': additional_lemma}))
+                                if (additional_tag, additional_lemma) not in seen:
+                                    tags.append(({'additional':True, 'tag': additional_tag, 'prob': -1, 'lemma': additional_lemma}))
+                                seen.add((additional_tag, additional_lemma))
                             else:
-                                additional_arcs.append(({'additional':True, 'tag': additional_tag, 'prob': -1, 'lemma': additional_lemma}))
+                                if (additional_tag, additional_lemma) not in seen:
+                                    additional_arcs.append(({'additional':True, 'tag': additional_tag, 'prob': -1, 'lemma': additional_lemma}))
+                                    seen.add((additional_tag, additional_lemma))
             else:
                 # There are words for which Freeling selected analysis should be ignored (no analysis discarded).
                 # In principle, there is also one tag for which it should be done if the word is in the first position:
@@ -160,5 +167,5 @@ def get_selected_tags(self, w, override_dicts):
                 add_tags = ADD_TAGS[a.get_tag()].split(',')
                 for mt in add_tags:
                     additional_arcs.append(({'additional': True, 'tag': mt.strip(), 'prob': -1, 'lemma': a.get_lemma()}))
-
         return tags, additional_arcs
+
diff --git a/util/treebanking-scripts/ace-process-all.sh b/util/treebanking-scripts/ace-process-all.sh
@@ -7,6 +7,6 @@ directory="$1"
 
 for profile in "$directory"/*; do
   echo $profile
-  delphin process --options="-y --yy-rules --max-chart-megabytes=24000 --max-unpack-megabytes=24000" -g ~/delphin/SRG/grammar/srg/ace/srg.dat --full-forest --select i-tokens "$profile"
+  delphin process --options="-1 -p -y --yy-rules --max-chart-megabytes=48000 --max-unpack-megabytes=56000" -g ~/delphin/SRG/grammar/srg/ace/srg.dat --full-forest --select i-tokens "$profile"
 done
 
diff --git a/util/treebanking-scripts/ace-process.sh b/util/treebanking-scripts/ace-process.sh
@@ -5,4 +5,4 @@
 
 profile="$1"
 
-delphin process --options="-y --yy-rules -1" -g ~/delphin/SRG/grammar/srg/ace/srg.dat --full-forest --select i-tokens ~/delphin/SRG/treebanks/dev/all/$profile
+delphin process --options="-1 -p -y --yy-rules --max-chart-megabytes=48000 --max-unpack-megabytes=56000" -g ~/delphin/SRG/grammar/srg/ace/srg.dat --full-forest --select i-tokens $profile
Original file line number	Diff line number	Diff line change
Expand Up		@@ -5,4 +5,4 @@

		profile="$1"

		delphin process --options="-y --yy-rules -1" -g ~/delphin/SRG/grammar/srg/ace/srg.dat --full-forest --select i-tokens ~/delphin/SRG/treebanks/dev/all/$profile
		delphin process --options="-1 -p -y --yy-rules --max-chart-megabytes=48000 --max-unpack-megabytes=56000" -g ~/delphin/SRG/grammar/srg/ace/srg.dat --full-forest --select i-tokens $profile